5bf77adef9
test(terminal): use managed temp directory
2026-07-29 13:29:52 +08:00
cec7f69f6c
test(terminal): enforce decoration regression coverage
2026-07-29 13:25:54 +08:00
12c8e03541
fix(terminal): speed dense keyword highlighting ( #2251 )
2026-07-29 12:26:07 +08:00
05f406ebab
fix(ai): use writable temp workspace for Cursor CLI MCP ( #2558 )
...
Packaged Netcatty launched from Finder often has process.cwd() === "/", so
merging .cursor/mcp.json failed silently and netcatty-remote-hosts never
mounted. Point CLI turns at Netcatty temp and fail loudly on merge errors.
Co-authored-by: Cursor <cursoragent@cursor.com >
2026-07-28 11:35:17 +08:00
e74859f786
fix(ci): proxy GitHub screenshots before research ( #2551 )
2026-07-27 23:15:25 +08:00
5a580ce923
feat(windows): add Explorer open-in-terminal menu
...
Closes #2513
2026-07-27 19:59:15 +08:00
b838a60e7c
Merge pull request #2539 from binaricat/codex/fix-cursor-research-output
...
fix(ci): accept fenced Cursor research status
2026-07-27 19:53:15 +08:00
709c7ec097
docs: add privacy and code signing policies
2026-07-27 19:45:44 +08:00
93f23b9312
fix(ci): require identical research candidates
2026-07-27 19:39:58 +08:00
7c90d40159
fix(ci): validate all research status candidates
2026-07-27 19:37:30 +08:00
2d3316093c
fix(ci): reject conflicting research statuses
2026-07-27 19:34:32 +08:00
0bb06dca72
fix(ci): prefer rebuilt final research fence
2026-07-27 19:31:20 +08:00
0d2c435687
fix(ci): parse Cursor partial stream boundaries
2026-07-27 19:28:10 +08:00
f2c47d2ab7
fix(ci): distinguish Cursor final flushes
2026-07-27 19:23:21 +08:00
d87029e200
fix(ci): trust valid terminal research result
2026-07-27 19:19:31 +08:00
25669b1118
fix(ci): rebuild final research envelope
2026-07-27 19:15:57 +08:00
c91aa04945
test(ci): handle split research fences
2026-07-27 19:13:17 +08:00
6dd3868a90
fix(ci): prefer final Cursor research status
2026-07-27 19:10:01 +08:00
b1221501c5
fix(ci): trust only final research message
2026-07-27 18:08:43 +08:00
28be1d396c
fix(ci): accept fenced Cursor research status
2026-07-27 18:06:45 +08:00
28ba165c3e
Merge pull request #2532 from binaricat/codex/fix-stuck-lint-test
...
fix(ci): prevent lint-and-test hangs
2026-07-27 16:50:41 +08:00
198300b062
fix(ci): prevent lint-and-test hangs
2026-07-27 15:49:37 +08:00
8473421b7a
Merge remote-tracking branch 'origin/main' into codex/pr-2413-finish
2026-07-27 14:56:21 +08:00
71e150504e
fix(ci): preserve trusted Cursor tool path
2026-07-27 14:43:17 +08:00
ae4400cdb8
fix(ci): run credential probe in agent mode
2026-07-27 14:35:49 +08:00
c33ae22965
fix(ci): isolate Cursor agent credentials
2026-07-27 14:32:51 +08:00
ad72ca9dee
Merge remote-tracking branch 'origin/main' into codex/pr-2413-finish
2026-07-27 13:55:58 +08:00
ac8ef5de38
fix(plugins): complete provider contract and lifecycle handling
2026-07-27 13:55:43 +08:00
fa5d1ec56c
fix(ci): pass Cursor auth to agent runs
2026-07-27 13:54:02 +08:00
3b75d9131c
fix(ci): execute Cursor sandbox probes directly
2026-07-27 12:11:32 +08:00
2cce98cf84
fix(ci): correct Cursor sandbox command parsing
2026-07-27 12:06:10 +08:00
27517f0d76
fix(ci): restore Cursor issue automation
2026-07-27 11:10:47 +08:00
2840463b3d
fix(automation): gate ready-state follow-ups and research web denials
...
Only automation bot PRs re-check source-issue pending comments before
mark-ready, so maintainer Fixes-only PRs are not stuck draft. Write
WebSearch/WebFetch denials after the isolated research pass so classify
and follow-up research can still call those tools.
2026-07-26 21:43:37 +08:00
b9fc1dfb7f
fix(plugins): address provider review gaps
2026-07-26 00:44:13 +08:00
768d5b9f23
merge: resolve main into continuous issue follow-ups
...
Keep PR follow-up routing (trigger_comment_id, pending issue gates) and
main's head_ref concurrency plus implement issue.json preparation.
2026-07-25 18:55:59 +08:00
5633631840
fix(ci): reduce flaky failures and PR runtime ( #2474 )
...
* fix(ci): reduce flaky failures and PR runtime
* fix(ci): avoid label scope failures
* fix(ci): tolerate GitHub download outages
* fix(ci): cover packaged inputs in PR triggers
* fix(ci): keep test-only changes out of package builds
* perf(ci): reuse unchanged ET PR builds
* fix: address Codex review on PR #2474
* fix(ci): harden automation PR handoffs
* fix(ci): preserve reused implementation branches
* fix(ci): deduplicate automation follow-ups
* fix(ci): deduplicate automation follow-ups
* fix(ci): ignore remapped stale Codex findings
* fix(ci): use the established handoff label
* fix(ci): isolate manual package validations
* fix(ci): refresh existing Nix metadata PRs
* fix(ci): hand off divergent implementation branches
* fix(ci): serialize Homebrew tap updates
---------
Co-authored-by: netcatty-bot <308658023+netcatty-bot@users.noreply.github.com >
2026-07-25 15:04:22 +08:00
e3b2c6a5e4
fix(terminal): preserve wakeups for pending TUI output ( #2470 )
...
Co-authored-by: bincxz <16399091+binaricat@users.noreply.github.com >
2026-07-25 12:10:47 +08:00
f88c66caf9
fix(automation): restore sandboxed Codex follow-ups
2026-07-25 11:56:56 +08:00
093c997cb5
fix(terminal): prevent Herdr rendering freezes ( #2472 )
...
* fix(terminal): prevent Herdr rendering freeze
* test(terminal): harden xterm overflow coverage
* test(terminal): stabilize WebGL overflow probe
* fix(terminal): keep visible xterm writes asynchronous
* fix(terminal): avoid interrupting xterm writes
* fix(terminal): reject incomplete output drains
* fix(terminal): keep failed snapshots atomic
* fix(terminal): cover Herdr rendering freezes
* fix(terminal): arbitrate pause leases in main
* test(terminal): keep Herdr timeout active
2026-07-25 08:39:59 +08:00
efecf436a7
feat(automation): add isolated web research pass
2026-07-25 00:47:42 +08:00
a55e9e7e7f
feat: continue issue automation after follow-ups
2026-07-25 00:02:49 +08:00
87b828c254
fix(terminal): complete Windows ConPTY clear handling ( #2465 )
...
* fix(terminal): enable ConPTY clear for Mosh and ET
* test(terminal): exercise Windows ConPTY spawn options
* fix(terminal): align ConPTY clear ABI
* fix(terminal): restore patched ConPTY runtime files
* fix(terminal): preserve patched ConPTY in packages
* test(ui): keep lazy-load comments ASCII-only
* fix(terminal): limit patched node-pty rebuild
* test(terminal): avoid platform-specific executable stubs
2026-07-24 22:58:38 +08:00
48b2cd1f67
fix(ci): close Cursor automation handoffs and label state machine ( #2461 )
...
- Format recovery dispatches triage before clearing invalid-format
- Shared CJK-friendly title rules; codex label terminal state machine
- Research steps in classify; agent TITLE + PR body for bot PRs
- Single @codex review via CODEX_REQUEST_GITHUB_TOKEN (binaricat)
2026-07-24 20:10:28 +08:00
3ff16920fd
fix(ci): make triage replies plain and drop Cursor disclaimer ( #2447 )
...
Public issue comments no longer show a generated-by line. Grounding stays
in code_findings/reasoning only so reply can use UI language without file
paths, symbols, or dense parentheses. Classify prompt and rewrite templates
match that tone.
2026-07-24 15:08:47 +08:00
27a42e3a76
fix(ci): Cursor triage crash + already_available how-to close ( #2439 )
...
* fix(ci): stop Cursor issue triage crashing on Search pagination
Octokit normalizes Search API pages so response.data is already the items
array. Mapping response.data.items returned undefined and then crashed on
candidate.user during daily-limit admission — so every valid new issue
failed before Cursor classify could run.
* fix(ci): classify already-available issues with how-to and auto-close
Add an already_available triage category for requests the product already
ships: after code-grounded inspection, reply with the concrete UI entry
path, then close as completed. Low confidence falls back to other instead
of closing. Keeps the Search pagination fix so new issues can classify.
2026-07-24 14:01:11 +08:00
31d4e16055
fix(ci): prefer local UI polish as feature_quick_win
...
Triage was biased toward feature_defer ("be conservative" + treating existing
UI tests / layout preference as strategic). Prefer quick wins for small,
same-panel UI work; reserve defer for multi-module or true product strategy.
Slightly lower the auto-downgrade floor for feature confidence to 0.7.
2026-07-24 10:12:05 +08:00
f160e9ad51
fix(ci): require code search and file reads before triage replies
...
Classification must open real source files first. The prompt mandates
search→read→classify, JSON now requires code_paths/code_findings, and
normalizeClassification rejects ungrounded replies so agents cannot answer
from issue text alone.
2026-07-24 10:04:06 +08:00
4d18ea811d
fix(ci): harden Cursor automation until Codex CLI review is clean
...
Address multi-round Codex CLI findings: trusted control markers, reaction-only
clean detection with head pins, draft-until-clean gates, secret leak scans
isolated from agent git hooks, fix eligibility limited to maintainers/actions
bot, P3 handoff, open-PR-only bot PR detection, and related safety fixes.
2026-07-24 04:27:59 +08:00
073083c6ee
fix(ci): address Codex review findings for automation safety
...
Isolate publish onto fresh runners, harden Codex loop head scoping,
protect fork bootstrapping, and polish classification replies and
mark-ready error handling after local codex review.
2026-07-24 01:54:30 +08:00
ba5932e285
fix(ci): raise default Codex fix loop cap to 40 rounds
...
Allow longer Cursor fix ↔ @codex review cycles on own/bot PRs; still
overridable via CURSOR_CODEX_FIX_MAX_ROUNDS.
2026-07-24 01:10:30 +08:00
bincxz